Limina crawls your sites, applications, and documents, tests them against WCAG 2.1 and 2.2, traces every finding back to the component that caused it, and tells your team what to fix first. Then it watches for the regressions.
The platform
Most teams run a site scanner, a separate document vendor, a spreadsheet of manual audit results, and a tracker that knows about none of them. Limina puts discovery, testing, remediation, monitoring, and reporting on the same record so the compliance story assembles itself.
Crawl, test, and continuously monitor websites and applications against WCAG 2.1 and 2.2.
Discover PDFs across your sites, remediate them, and validate the result at repository scale.
ships as PercevaWCAG 2.0 through 2.2 by criterion, mapped to ADA, Section 508, EN 301 549, PDF/UA, and more.
Assign, fix, verify, and close, with the right issue reaching the right role.
Root cause analysis, prioritization, AI assisted fixes, and a copilot that answers in plain language.
Scores, trends, portfolio rollups, executive reporting, and audit ready evidence.
Web accessibility
Limina crawls public sites, portals, and authenticated applications on the schedule you set, scores every page, and compares each run against the last. Findings arrive ranked by severity and impact rather than in one undifferentiated list.
A number that says twenty thousand errors tells a team nothing they can act on. What matters is what moved since the last deployment, and whether anything got worse.
| New issues introduced | 146 | 27 pages |
|---|---|---|
| Regressions on previously fixed items | 37 | needs triage |
| Issues resolved | 1,204 | verified |
| Documents added by crawl | 88 | queued |
| Accessibility score | 87 | up 5 this quarter |
Testing after publication means fixing in production. Limina runs inside the pipeline, so a build that introduces a critical violation can be stopped before it ships.
Document accessibility
Most organizations underestimate their document exposure by an order of magnitude, because nobody has ever inventoried it. Limina crawls for linked documents while it crawls for pages, so the web and document pictures come from the same scan.
The document engine ships as Perceva, and is sold on its own for organizations whose exposure sits entirely in PDFs. Inside Limina it is one module of six, reporting into the same score and the same evidence record.
| File | Pages | Published | Status | Complexity | Owner |
|---|---|---|---|---|---|
| fy26-budget-narrative.pdf | 218 | 2026-02-11 | 14 violations | high | Finance |
| permit-application-b.pdf | 6 | 2025-11-30 | forms review | medium | Permitting |
| council-minutes-2026-08.pdf | 42 | 2026-08-19 | PDF/UA pass | low | Clerk |
| deed-scan-04117.tif.pdf | 3 | 1998-06-02 | recognizing text | high | Records |
| transit-fare-schedule.pdf | 11 | 2026-07-04 | tables untagged | medium | Transit |
| vendor-brochure-2019.pdf | 8 | 2019-03-15 | archive proposed | n/a | Procurement |
Tagging and structure, logical reading order across multi column layouts, alt text with reviewer approval, table header association including merged and multi level headers, form labels and tab order, optical character recognition for scans, and document properties including language, title, bookmarks, and navigation.
Uncertain items are routed to human review rather than passed through. Complex tables, poor scans, and unlabeled forms are the usual exceptions, and they stay in the queue until a person resolves them.
Compliance
Automated testing cannot establish full WCAG conformance on its own, and any tool that claims otherwise is selling you a liability. Limina combines automated results with guided manual checks and records a verdict against every success criterion.
Specialists work through structured checks with the criterion, the method, and the expected result in front of them. Each verdict carries notes and evidence, and rolls up into the same conformance record as the automated results.
Testing happens against WCAG. Reporting happens against whatever your obligation is called in your jurisdiction, without running the work twice.
Workflow
Every detection becomes a tracked item with a location, a standard, an affected user group, and an owner. Developers get code. Editors get content. Remediators get documents. Nobody gets a spreadsheet of everything.
Verification re-tests the specific criterion on the affected pages. An issue cannot be closed on assertion alone.
Accessibility work fails when it lands on one person. Limina routes each class of issue to the role that can resolve it, and gives each role a view scoped to their own queue.
Editors see accessibility feedback inside the tools they already write in, and designers see contrast, focus states, and touch target problems while the component is still a design.
Intelligence
Scanners count occurrences. That is why a mid sized site produces twenty thousand findings and a team that gives up in week two. Limina groups findings by template, component, CMS block, and design system element, so the work is sized by cause rather than by symptom.
Each root cause is scored on severity, user impact, page traffic, frequency, legal risk, and how many occurrences a single fix clears. What comes back is a ranked worklist, not a backlog.
Not just what failed. What it breaks, which criterion it maps to, who it affects, what the code does now, and what it should do instead.
current code
<button class="btn-pay" onClick={submit}> <svg class="icon-lock" /> </button>
suggested fix, pending developer approval
<button class="btn-pay" onClick={submit} aria-label="Submit payment"> <svg class="icon-lock" aria-hidden="true" /> </button>
The copilot reads your scan history, your issue records, and your document queue, and answers the questions people actually ask in a status meeting.
why did our accessibility score drop this week
The score moved from 91 to 87 after the 2 September deployment. 146 new errors were introduced across 27 pages, and 112 of them originate from the redeployed global navigation component.
The navigation change removed the accessible name from the menu toggle and altered focus order on mobile viewports. Fixing that one component would recover an estimated 3.4 of the 4 points lost.
what should we fix first
Seventeen root causes account for 18,490 of your 24,806 open occurrences. The first three are worth prioritizing this sprint.
are we ready for the title II deadline
Not yet, on current velocity. Web properties are at 89% of WCAG 2.2 AA criteria passing, with 6 failures and 5 items still needing manual review. Documents are the constraint: 2,194 files still require remediation and 1,973 are in processing.
At the current remediation rate of 7,481 items a month, the document backlog clears roughly six weeks after your date. Raising document throughput or approving the 1,140 proposed archive candidates would close that gap.
Analytics
Executives get accessibility risk, trend, and remediation velocity. Specialists drill from that same number down to the element that caused it, without switching tools or reconciling two sets of figures.
87/100
up 5 points this quarter| Websites | 91 | |
|---|---|---|
| Web applications | 84 | |
| PDFs and documents | 76 | |
| Forms | 89 | |
| Content | 93 |
Organization to department to property to page to component to issue, and the same path through document libraries down to the element on the page.
If a complaint arrives, the question is not whether you are perfect. It is whether you can show what you knew, what you did, and when. Limina keeps that record continuously rather than assembling it under pressure.
Accessibility work that lives in its own tool gets ignored. Limina pushes issues into the trackers your teams already open every morning.
No, and be careful with any vendor that says it does. A meaningful share of WCAG requires human judgement: whether alt text is actually meaningful, whether an error message is understandable, whether a custom widget behaves the way its role implies. Limina automates detection, automates the fixes that can be safely automated, proposes the rest for developer approval, and structures the manual testing that has to happen either way.
Never. Suggested code is presented alongside the current code, the criterion, and the reason. A developer reviews and approves before anything reaches a branch. The same applies to documents: generated alt text is drafted, then approved, edited, or replaced by a reviewer.
A scanner produces a list of occurrences. Limina traces those occurrences to the shared components and templates that generate them, scores them by severity, traffic, and fix leverage, and gives you a ranked worklist. It also covers the documents linked from your site, which scanners generally do not, and keeps the compliance evidence that scanners do not produce.
Yes. Authenticated crawling is supported, and application flows can be tested in the pipeline as part of your existing automated suites rather than only from the outside.
Not necessarily. The discovery, inventory, and triage stages are useful on their own, and many organizations use them to size the work and route it: automated remediation for the bulk, a specialist vendor for the genuinely complex material. Either way the status comes back to one record.
A first crawl of a mid sized property, including its documents, typically produces a full inventory and a prioritized worklist without any configuration beyond pointing it at the domain. Deeper value comes from the second scan onward, when change and regression tracking start.